Washington EO 24-01 (Artificial Intelligence)

EXECUTIVE ORDER 24-01 ARTIFICIAL INTELLIGENCE WHEREAS, generative Artificial Intelligence (AI) represents a significant leap forward in technology, by generating novel text, images, and other content, which has the capacity to transform the way that the State conducts business and serves the public; and WHEREAS, generative AI has the potential to catalyze innovation and the rapid development of a wide range of benefits for Washingtonians; and WHEREAS, generative AI can enhance human potential and creativity but must be deployed and regulated carefully to mitigate and guard against a new generation of risks, harms, and perpetuation of existing inequities; and WHEREAS, the state of Washington is committed to accuracy, reliability, and ethical outcomes when adopting generative AI technology, engaging and supporting historically vulnerable and marginalized communities that are also denied a voice in the data collected about them and how it is used, and serving all those who reside, work, or do business in Washington in a transparent, engaged, and equitable way; and WHEREAS, the state of Washington seeks to realize the potential benefits of generative AI for the good of all those who call Washington home, through the development and deployment of generative AI tools that improve the equitable and timely delivery of services, while balancing the benefits and risks of these new technologies; and WHEREAS, the Washington state workforce is vital to Washington's continued prosperity and the state seeks to harness the potential of generative AI in an ethical and equitable way for the benefit of the state government workforce; and WHEREAS, the exponential growth of generative AI technologies should be documented, and the potential risks of their use, including but not limited to privacy, cyberattacks, disinformation, deception, and discrimination or bias, should be acknowledged, planned for, and mitigated; and WHEREAS, Executive Branch agencies across state government have been engaging, and will continue to engage, with communities and other industry stakeholders to explore strategies to promote the safe, responsible, ethical, and efficient development and use of generative AI technologies in Washington; and WHEREAS, Washington state should promote the adoption of generative AI technologies in a manner that aligns and is consistent with Washington state’s policies, goals, and values for equity in government employment, services, education, and contracting as well as for all people who call Washington home.

NOW, THEREFORE, I, Jay Inslee, Governor of the state of Washington, by the power vested in me by the Constitution and the statutes of the state of Washington, do hereby order and direct all executive and small cabinet agencies as follows: 1. It is the goal of the state to create a transparent and consistent environment to use generative AI technology and to facilitate the adoption of generative AI technology by the State in a manner that balances the benefits and risks to Washingtonians. For the purposes of this Order: a. “Generative AI Technology” is a technology that can create content, including text, images, audio, or video, when prompted by a user. Generative AI systems learn patterns and relationships from large amounts of data, which enables systems to generate new content that may be similar, but not identical, to the underlying training data. b. “High-Risk Generative AI System” means systems using generative AI technology that creates a high risk to natural persons' health and safety or fundamental rights. Examples include biometric identification, critical infrastructure, employment, health care, law enforcement, and administration of democratic processes. 2. By September 2024, Consolidated Technology Services, also known as Washington Technology Solutions (WaTech), in collaboration with cabinet agencies, shall submit a report to the Governor that identifies potential generative AI initiatives pertinent to agency operations and objectives. The report shall assess the feasibility, benefits, and challenges of integrating generative AI into agency operations and services. The report must also include a plan and timeline for WaTech to establish the infrastructure to conduct pilots of projects involving generative AI technology, including approved environments to test such projects.

3. By September 2024, WaTech, in collaboration with the Department of Enterprise Services (DES), shall issue initial guidelines for public sector procurement, uses, and ongoing monitoring of the use of generative AI technology. The guidelines should build on guidance from the White House's Blueprint for an Al Bill of Rights and the National Institute for Standards and Technology's Al Risk Management Framework and address safety and effectiveness, algorithmic discrimination, data privacy, and cybersecurity. These guidelines shall include a requirement that any vendor providing a High-Risk Generative AI System to an agency certify that the vendor has implemented an AI governance program consistent with the National Institute of Standards and Technology Artificial Intelligence Risk Management Framework. WaTech and DES shall make the guidelines publicly available via posting on relevant agencies’ websites. 4. No later than January 2025, DES, in collaboration with WaTech and the Office of Equity, shall submit a training plan for state government workers on the use of generative AI technology. The training plan shall address, at minimum, the potential benefits to state operations and constituent services, risks to citizens related to unfair and discriminatory outcomes, privacy and cybersecurity risks, and other potential harms including automation bias.

5. By December 2024, WaTech, in consultation with the Office of Equity, community members, Tribal governments, subject matter experts, and other relevant state agencies, shall develop guidelines for agencies to analyze the impact that adopting generative AI technology may have on vulnerable communities, including criteria to evaluate equitable outcomes in deployment and implementation of high-risk use cases. These guidelines and criteria shall inform whether and how an agency deploys a particular AI System. WaTech shall periodically revise the guidelines and criteria, when appropriate, and establish a consultative process with academia, industry experts, and representatives from advocacy organizations that represent communities that are disproportionately vulnerable to being harmed by algorithmic bias. WaTech and the Office of Equity shall make the guidelines publicly available by posting on their respective agency’s website. 6. By January 2025, DES, in collaboration with WaTech, shall update the state's technology procurement and contract term templates, incorporating equity, bias, and algorithmic risk assessments as well as analysis and feedback obtained from the initial guidelines issued in (3) above and other relevant, community members, academic and industry experts, and the state government workforce or organizations from (5).

7. To ensure algorithmic justice as it relates to digital equity, by September 2024, the Office of Equity will oversee the development, implementation, and ongoing support of an overarching framework for accountability in the fair, ethical, and equitable use of generative AI by the state. a. The accountability framework will be created in partnership with WaTech, DES, and Workforce Training and Education Coordinating Board to de-silo and remove barriers for agencies seeking to test and implement generative AI tools. b. The accountability framework must be comprised of a transparent, systematic, and continuous monitoring of outcomes and impacts on communities and must include the intentional and consistent engagement of communities. c. The accountability framework should be informed by subject-matter experts from across the Enterprise, community members, Tribal governments, impacted front-line state employees, and academic and industry stakeholders whose purpose will be to co-create an assessment process and performance measures.

8. By December 2024, WaTech will produce guidance on the risk assessments for the deployment of High-Risk Generative AI Systems. Assessments must leverage existing security and privacy assessment processes, and must include the following: a. Information about the High-Risk Generative AI System, including whether the HighRisk Generative AI System is provided by a third party, the name and address of the third party, and relevant state agency; b. The intended uses of the High-Risk Generative AI System; c. Assessment of the fitness of the High-Risk AI System for the intended purpose; d. Assessment of impacted communities, benefits, harms, and mitigations of the HighRisk AI System; e. An evaluation of the potential harms of the High-Risk Generative AI System which may include harms to individuals and groups, discriminatory or unfair outcomes, deceptive practices, societal risks, privacy and cybersecurity considerations, and national security concerns; f. An assessment of mitigations including but not limited to consideration of restricted uses and limitations on use, policies, deidentified data, and commercial terms; and g. Information about the agency approach to generative AI governance that is consistent with the AI Risk Management framework published by the National Institute of Science and Technology. 9. By December 2024, the Office of Financial Management (State Human Resources Division) in collaboration with WaTech, Washington’s Workforce Training and Education Coordinating Board and organizations that represent state government employees, will assess the impact of generative AI on the state workforce, develop strategies to mitigate any negative impacts, and support programs that help workers develop the skills and knowledge they need to successfully use generative AI and report their findings to the Governor.

10. By January 2025, members of Washington’s Workforce Training and Education Coordinating Board must identify and create research opportunities and partnerships with research institutes at a state and federal level to power innovation and expansion of generative AI technology and education. The goals will be to expose students and researchers to emerging opportunities regarding the development of generative AI Systems and accelerate training and AI workforce development. The members are encouraged to collaborate and identify opportunities to: a. Build workforce pathways that ensure a pipeline of talent to the industry that is both equitable and robust; b. Develop training programs geared towards generative AI workforce development; c. Generate basic and applied research to continue leading future generations of generative AI technologies; and d. Engage national organizations, such as the National AI Research Resource (NAIRR) Taskforce, to achieve the goals stated above.

11. Agencies subject to my authority shall implement the directives in this section consistent with current federal and state laws. In anticipation of a harmonized state-federal legal landscape, any provisions in this Order inconsistent with federal law shall be superseded by federal law. 12. This Order is not intended to, and does not, create any legal right, entitlement or benefit, substantive or procedural, enforceable at law or in equity, against the state, its agencies, departments, entities, officers, employees, or any other person. This order takes effect immediately. Signed and sealed with the official seal of the state of Washington on this 30th day of January, AD, Two Thousand and Twenty-Four, at Olympia, Washington.